Special Seminar
Friday, March 3, 2023
11:00 a.m.
Room 2108 Chemical and Nuclear Engineering Building
For More Information:
Emmanuel Duh
301 405 1935
eduh@umd.edu

Title:Building predictive macroscale models from microscale information: A combined physics andmachine learning approach

Abstract: Ensuring a sustainable future requires the development of new chemistries and materials. In order to accelerate materials discovery, new computational methods are needed to identify suitable candidates for experimentation and identify important structural characteristics of synthesized materials. All of this must be done in the context of limited datasets for materials, properties, and conditions of interest. In this work, we combine data- and physics-based methods to develop computational models that can accurately predict macroscale properties from primarily microscale information. Applications include faster and more accurate kinetic models for structure sensitive reactions, characterization of complex microstructures from spectroscopy, and transfer learning approaches for application of machine learning to small chemical datasets.

Bio: Joshua Lansford is an ARL Postdoctoral Fellow at The Massachusetts Institute of Technology (MIT) in Klavs Jensen’s group, where he is designing transfer learning techniques to translate machine learning models to small chemical datasets. Prior to joining MIT, Josh was a Blue Waters Graduate Research Fellow at the University of Delaware, where he obtained his Ph.D. in chemical engineering under the advisement of Dion Vlachos and developed computational methodologies for catalyst design. Dr. Lansford is the recipient of the North American Catalysis Society (NACS) Kokes Award, the Catalysis Club of Philadelphia (CCP) Ted Koch Travel Award, and the Phillip and Ruth Evans Fellowship. He has also been awarded travel grants for the Berkeley Deep Learning for Science School, the International Symposium on Chemical Reaction Engineering (ISCRE25), and the American Institute of Chemical Engineering (AIChE) Annual Meeting. His current interests lie in developing new computational tools for sustainable chemistries with applications in direct air carbon capture and upgrading, microplasticremediation, fuel cells, and batteries.

This Event is For: Graduate • Faculty